The Possible Role of Cytokeratin 8 in Cadmium-Induced Adaptation and Carcinogenesis Free

Chronic exposures to cadmium derivatives are nearly unavoidable in daily life, such as from food, soil, water, and airborne particles. Although it has been well documented that cadmium exposure can cause a wide variety of adverse health effects and cancer development, the mechanism of its action has not been well established (1–4). It has been suggested that cadmium exerts genotoxic or carcinogenic effects through an alteration in expression of immediate early genes (c-jun, c-fos, and c-myc) or the tumor suppressor gene p53 (5–8). However, among all its effects, depressed apoptosis observed during cadmium adaptation is one of the important mechanisms driving cadmium carcinogenesis; this is achieved by selection of adapted cells in which apoptotic response is significantly attenuated (9, 10).

We previously reported that cadmium-adapted alveolar epithelial cells are protected from oxidant-induced apoptosis and along with the induction of stress gene expression (metallothionein, glutathione S-transferase, and γ-glutamylcysteine synthetase; ref. 11). In addition, a more recent finding from our group showed that these adapted cells develop tolerance to cell death generally due to perturbation of the c-Jun NH₂-terminal kinase (JNK) signaling pathway and the nonresponsiveness of JNK phosphorylation is critical for the cadmium tolerance in these cells (9). However, the mechanism(s) of how JNK inactivation is still unknown or whether the presence of other yet unknown important regulatory mechanisms governs the death tolerance. To gain further insights into the mechanism for cellular long-term survival in response to cadmium, here, we did comparative proteome analysis of these cells with parental cells that were either untreated or treated with cadmium. Protein spots altered in cadmium-adapted cells were identified by peptide mass fingerprinting using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) and database searching. Remarkably, cytokeratin 8 (CK8) and cytokeratin 14 (CK14) were increased only in the adapted cells but not in parental cells treated with cadmium, suggesting that these proteins were secondary cadmium-responsive proteins.

Cytokeratin proteins constitute the largest and most complex class of intermediate filaments. They are expressed in epithelial cells throughout the body where they form structural networks to span the cell cytoplasm linking the plasma membrane, nucleus, and other cytoskeletal components. Keratin filaments are obligate heteropolymers consisting of type I and II proteins in a 1:1 ratio. The 12 type I and the 8 type II keratin monomers pair in certain defined combinations in a tissue-specific and developmentally regulated manner (12). The expression of certain keratin proteins in a cell at a given time is defined as its state of differentiation, and modulation of differentiation program has been well documented in cell transformation and cancer development (12). Recently, cytokeratin 19 has been found to be a unique biomarker to judge the invasiveness of breast cancer cells (13). Although modulation of differentiation program has been known for many years and used to distinguish various cell types or as biomarkers in cancer progression, very few studies have addressed the underlining importance of this transition. In this study, we provided evidence to show that increased expression of cytokeratins in cadmium-adapted cells represents an additional survival mechanism that resists cadmium-induced apoptosis and it is unprecedented in our study that cells respond to long-term cadmium exposure by modulating keratin dynamics.

Materials. Cadmium chloride (CdCl₂) was purchased from Sigma (St. Louis, MO). PlusOne 2-D Clean-Up kit and Silver Staining kit were purchased from Amersham Biosciences (Uppsala, Sweden). All other general chemicals were purchased from Amersham Biosciences and Sigma. Antibodies used for Western blot were purchased from Sigma, Upstate Biotechnology (Lake Placid, NY), and Santa Cruz Biotechnology (Santa Cruz, CA).

Cell culture and transfection. A rat lung epithelial cell line (LEC) was isolated and characterized by Li et al. (14). This cell line is designated as Cd^S in this study for its cadmium sensitivity in comparison with cadmium-resistant (Cd^R) cell lines developed in our laboratory. Development of Cd^R cells has been previously described by Lau et al. (9). In all experiments, cells were subjected to treatment at ∼75% confluence. Cell viability was measured by naphthol blue black (NBB) staining assay as described previously (15).

LEC cells were cultured in six-well plates before transfection with 100 nmol/L small interfering RNA (siRNA) duplexes against CK8 using LipofectAMINE 2000 (Invitrogen, Carlsbad, CA). At 48 h after transfection, the cells were harvested for two-dimensional PAGE analysis or Western blot analysis after cadmium treatment. The sequence of the region targeted by CK8 siRNA (Qiagen, Lafayette, CO) 5′-CAGCAUCAUUGCUGAGGUCAA-3′ was nucleotides 771 to 791 of the CK8 open reading frame. Control cells were incubated with the transfection reagent for as long as the transfected cells. Control experiments were carried out with the nonspecific siRNA control IX containing 47% GC (Dharmacon Research, Lafayette, CO) under the conditions of CK8 siRNA or with LipofectAMINE 2000 alone. As there were no differences in the CK8 level between the control siRNA and LipofectAMINE 2000 alone, we commonly used the transfection agent alone for control experiments.

Cell lysate preparation and conditions of Western blot and two-dimensional PAGE. After treatment, cells were then washed thrice with ice-cold PBS, scraped into centrifuge tube, and then harvested by centrifugation at 1,000 × g for 5 min at 4°C.

For Western blot analysis, cell pellets were lysed in radioimmunoprecipitation assay buffer according to the protocol described by Upstate Biotechnology. Equal amounts of proteins (40 μg) were fractionated on a SDS-polyacrylamide gel and transferred onto polyvinylidene difluoride membranes. The membranes were blocked with 5% nonfat dry milk in PBS containing 0.05% Tween 20 and probed with various primary antibodies. After incubation with secondary antibodies, immunoblots were visualized with the enhanced chemiluminescence detection kit (Amersham Biosciences).

For two-dimensional PAGE analysis, cell pellets were lysed in lysis buffer [8 mol/L urea, 4% (w/v) CHAPS], incubated on ice for 30 min, and centrifuged at 16,000 × g for 5 min at 4°C. The supernatant was saved and then further purified by using the PlusOne 2-D Clean-Up kit in accordance with the manufacturer. The purified samples were finally redissolved in rehydration buffer (8 mol/L urea, 2% CHAPS), aliquoted into several tubes, and stored at −80°C after protein quantitation. Two-dimensional PAGE was done on 80 μg of cleaned-up cell extract with Amersham Biosciences IPGphor IEF and Hoefer SE 600 electrophoresis units. All gels were visualized by silver staining using the PlusOne Silver Staining kit in accordance with the manufacturer.

Image analysis, MALDI-TOF-MS analysis, and protein identification. The stained gels were scanned using an ImageScanner (Amersham Biosciences) operated by the LabScan 3.00 software. Image analysis was carried out by using the ImageMaster 2D Elite software 4.01. Only up-regulated/down-regulated spots (more than ±2-fold) or spots that either appeared/disappeared were selected for analysis with MS. Protein spots were excised and transferred into siliconized 1.5 mL Eppendorf tubes. Gel chips were destained, dehydrated with acetonitrile, and then rehydrated in trypsin solution (10 μg/mL in 25 mmol/L NH₄HCO₃) at 37°C overnight. The digest was then applied onto a sample plate and coated with matrix (α-cyano-4-hydroxycinnamic acid). Tryptic peptide MS were obtained using a Voyager-DE STR MALDI-TOF-MS (Applied Biosystems, Foster City, CA). Protein identification was done by searching in NCBInr protein database using MS-Fit.¹

Statistical analysis. Statistical analysis was done by using two-tailed Student's t test, and P < 0.05 was considered significant. Data are expressed as the mean ± SD of triplicate samples, and the reproducibility was confirmed in three separate experiments.

Cadmium induces mitogen-activated protein kinase phosphatase-1 expression in cadmium-sensitive and cadmium-adapted cells. From our previous studies, we established cadmium-adapted LECs that exhibit resistance to cadmium-induced apoptosis (9, 11). In conjunction with primary response/defense proteins, we have shown that cadmium resistance in LECs is generally due to the perturbation of the JNK pathway (9). First, we suspected that desensitization of JNK in cadmium-adapted cells was due to the up-regulation of mitogen-activated protein kinase (MAPK) phosphatases (MKP), which dephosphorylates activated JNK, because the MKPs are responsible for inactivation of MAPKs in various cell types (16, 17). Among the MKPs, MKP-1 is the best characterized member of this family and has been shown to inactivate JNK. Various stresses that activate JNK can induce MKP (18, 19). Thus, the duration of JNK activation can be regulated by MKP through a feedback mechanism.

Elevated levels of MKP-1 have been shown in renal cell carcinoma cell lines in which they developed apoptotic resistance to subsequent drug treatment (20). To determine whether cadmium exposure would affect the MKP-1, we used antibody against MKP-1 to examine the effect of cadmium on Cd^S, Cd^R, and Cd^R cultured in the absence of cadmium for two passages [Cd^R(−2)] by Western blot throughout a 24-h time interval. Surprisingly, cadmium-adapted cells behaved similarly like cadmium-sensitive cells, which have undetectable basal levels of MKP-1. However, on 20 μmol/L CdCl₂ treatment, MKP-1 was rapidly induced only in Cd^S as indicated by the increase in MKP-1 expression (Fig. 1A). In contrast to Cd^S, the MKP-1 level on Cd^R and Cd^R(−2) was weakly stimulated or almost unchanged after exposure to 20 μmol/L CdCl₂. This is out of our speculations because Cd^R cells should at least have higher basal levels of MKP-1 to desensitize JNK. To see how cadmium-adapted cells respond to higher cadmium dosages on MKP-1 induction, cells were treated with higher concentrations of cadmium (50, 100, or 200 μmol/L) for 2 h and the levels of MKP-1 were determined (Fig. 1B). We chose a 2-h treatment because, at this time point, cadmium induced MKP-1 and activated MAPKs in preliminary experiments. Interestingly, the level of MKP-1 was markedly decreased in all the Cd^R cells compared with cadmium-sensitive cells at each level of cadmium treatment (at least in 50 and 100 μmol/L). Because Cd^R has higher basal level of metallothionein-1 mRNA than Cd^R(−2), which resulted in more efficient metal-binding capacity (9), the level of MKP-1 was also markedly decreased in Cd^R even exposed up to 200 μmol/L CdCl₂ compared with Cd^R(−2). Treatment of cells with 50 μmol/L CdCl₂ for 0.5-, 1-, and 2-h time intervals also showed similar results (Fig. 1C). Thus, JNK pathway desensitization in Cd^R was not due to the up-regulation of MKP-1 expression, as there was indeed less amount of MKP-1 induced in Cd^R cells. These results were indicative of nonessential role of MKP-1 in cadmium adaptation in both Cd^R and Cd^R(−). Therefore, this prompted us to further study the resistant factors by proteomic approaches.

Proteome profiles between cadmium-sensitive and cadmium-adapted cells. To search for proteins associated with cadmium adaptation, proteome profile of parental cells was compared with that of cadmium-adapted cells by using two-dimensional PAGE. Figure 2A depicted the silver-stained two-dimensional gel protein expression profile obtained after separation of protein lysates from these cells. Spot volume comparison was made between the samples with the ImageMaster program. In general, there were no significant differences between Cd^S and Cd^R, except for two spots. These two spots did not change in cadmium-treated parental cells, indicating their potential roles as other cadmium-responsive proteins that confer resistance to cadmium-induced apoptosis. Figure 2B was a comparative montage view of the regions shown in Fig. 2A, where significant differences in protein expression level were marked. These protein spots were excised and subjected to trypsin digestion, MALDI-TOF-MS measurements, and database searching with MS-Fit. These two proteins were identified as CK8 and CK14 (Tables 1 and 2).

CK8 and CK14 overexpressions in cadmium-adapted cells result from long-term cadmium adaptation. In the proteome data, the levels of CK8 and CK14 were increased only in adapted cells but not in cadmium-treated parental cells, suggesting that their up-regulation may be the secondary consequence of cellular response to cadmium. CK8 and CK14 are structural proteins belonging to the intermediate filaments. In many cancer cells, including human breast carcinoma, lung squamous cell carcinoma, and gastric adenocarcinoma (21–23), CK8 and CK14 are highly elevated. In addition, CK8 and CK14 have been shown to suppress apoptosis induced by a variety of agents, and the death inhibitory action of CK8 has been attributed to its ability to provide resistance to tumor necrosis factor (TNF)-induced apoptosis (24). Similarly, CK14 may function as an inhibitor of TNF-TNF receptor (TNFR) 1 signaling through an association with TNFR1-associated death domain protein (TRADD; ref. 25). Thus, we predicted that elevated CK8 and CK14 in the cadmium-adapted LECs is a part of the cellular responses to resist cadmium-induced apoptosis.

In parallel, we also did proteome analysis on cadmium-adapted cells after removal of cadmium in medium for various number of passages [i.e., Cd^R(−2) and Cd^R(−7); Fig. 2B]. The elevated cytokeratin expressions were stably maintained even after culturing the adapted cells in the absence of cadmium for at least seven passages (28 days; Table 1). Thus, the increased CK8 and CK14 expressions were consequences of long-term adaptation to cadmium but not inducible by short-term treatment. Because the data of CK14 in National Center for Biotechnology Information (NCBI) are a partial sequence from subtractive cloning, therefore, in this study, we first only focused our attention on CK8 to address its importance in cadmium adaptation and acquired tolerance.

The effects of CK8 gene silencing on cadmium sensitivity. From our proteomic studies, we suggest that if increased CK8 was involved in enhanced cadmium resistance, then removal of CK8 expression should block this protection. For this reason, we resolved to use loss-of-function method to silence the highly expressed CK8 in Cd^R by siRNA technique. To test the effect of silencing of the CK8 gene on cadmium sensitivity, cadmium-adapted cells were transfected with 100 nmol/L siRNA and grown for 2 days. The samples were processed for two-dimensional gel to determine the expression of CK8. As expected, transfection of cadmium-adapted cells with siRNA resulted in ∼60% reduction in CK8 gene expression (Fig. 3A). More importantly, the silencing of CK8 before the addition of cadmium reactivated JNK and significantly decreased the cell viability of cadmium-adapted cells compared with control cells (Fig. 3B and C). These results clearly showed that silencing of CK8 recurred cadmium-sensitive phenotype in the cadmium-adapted cells. To sum up, based on the results obtained from silencing of CK8 in cadmium-adapted cells by siRNA, we concluded that the elevated expression of CK8 in Cd^R LEC cells is crucial in protection against cadmium-induced apoptosis.

Chronic exposure to cadmium is associated with increased risk of cancers (1, 2). Our current knowledge with respect to the actions of cadmium has been gathered from numerous epidemiologic and clinical studies encompassing geographically diverse populations and multiple exposure scenarios. Despite the availability of data, mechanisms governing the carcinogenic effects have yet to be well established.

We resolved to use proteomic approach to identify the responsive proteins that are involved in cadmium adaptation. Unlike the primary cadmium-responsive proteins (stress proteins), CK8 and CK14 were up-regulated only in cadmium-adapted cells but not in cadmium-treated parental cells, suggesting that the up-regulation was due to the secondary consequence of cellular response to cadmium. Their up-regulation was not due to the presence of cadmium in the medium, as this also occurred in cadmium-adapted cells that were cultured in the absence of cadmium for as long as seven passages. These results were indicative of cytokeratin overexpressions in both Cd^R and Cd^R(−) once resistance to apoptosis develops. More importantly, we showed that siRNA-specific knockdown of CK8 in cadmium-adapted cells reactivated JNK and attenuated the cadmium resistance, indicating the potential role of CK8 in cadmium resistance.

CK8 and CK18 formed the major structural keratins in noncutaneous epithelial cells in liver, bowel, and single-layered epithelia. A recent study provided evidence for the key role of CK8/cytokeratin 18 (CK18) in moderating the signaling of TNF (24). It was shown that decreasing levels of CK8 and CK18 in cultured epithelial cells increased cellular sensitivity to killing by TNF due to the loss of binding of both CK8 and CK18 with the cytoplasmic domain of TNFR2. In addition, CK8 and CK18 diminished the TNF-dependent activation of JNK and the nuclear factor-κB transcription factor. In vivo, CK8- or CK18-null mice were shown to be more sensitive to TNF-mediated apoptotic liver damage (24). In a two-hybrid screen, TRADD specifically bound CK18 and CK14, type I keratins, but did not bind type II keratins, including cytokeratin 5 and CK8, or type III intermediate filaments, including vimentin, glial fibrillary acidic protein, and desmin. Because CK14 is preferentially expressed in the basal layer of epidermis, and in the same fashion as CK18, it is possible that epidermal CK14 may function as an inhibitor of TNF-TNFR1 signaling through an association with TRADD (25). In this study, although we have not further studied the function of CK14, from the literature and our results presented here, we speculate that it is also a potential cadmium-responsive protein that warrants future studies.

It has been shown that the phosphorylation of human CK8 at Ser⁷³ could involve a kinase from the MAPK family due to the serine-proline motif of CK8-Ser⁷³ phosphoacceptor site (26). By comparing the sequence surrounding the Ser⁶³ and Ser⁷³ phosphoacceptor sites in c-Jun with that surrounding position 73 in rat CK8, there is no such serine residue on this position. Hence, rat CK8 is not subjected to phosphorylation at Ser⁷³. However, similar consensus sequence was found at its Ser⁴³² position (tail domain; i.e., L-T-S⁴³²-P-G). Phosphorylation at Ser²³ and Ser⁴³¹ on human CK8 implicated that both the head and tail domains of CK8 can be phosphorylated as shown previously for cytokeratin 1, another type II keratin (27, 28). Hence, the Ser⁴³² position on rat CK8 is a putative JNK phosphorylation site. In addition, phosphorylation of CK8 on Ser⁴³¹/Ser⁷³ has been shown to protect it from ubiquitination (29). Thus, JNK may be involved in regulating CK8 ubiquitination by phosphorylating CK8. However, besides bearing the phosphoacceptor site, effective phosphorylation by JNK is facilitated by the presence of a specific JNK docking site (30, 31). To see if there is/are any possible JNK docking site(s) in rat CK8, we compared the docking site sequences of c-Jun and JunB with that of CK8. It was found that there are two short sequences in CK8 that resembled these JNK docking sequences (Table 3). It is possible that, in some cancer cells, CK8/CK18 duplexes may act as a JNK sequestering complex, thereby causing cytoplasmic retention of JNK from doing its nuclear tasks. Our results showed that CK8 is stably up-regulated in cadmium-adapted cell, which may correlate to the less JNK activation in these cells and result in apoptotic resistance. However, the hypothesis that the docking of JNK with CK8 would affect JNK shuttling in our studying model (LECs) will need to be further investigated.

In conclusion, we postulate that up-regulation of CK8 and CK14 may be responsible for apoptotic resistance in cadmium-adapted cells. Because CK8 and CK14 are overexpressed in various cancer cells, it is logical to assume that the elevation of these proteins could promote neoplastic cell survival as well as subsequent clonal expansion and then progression of tumor development. On the other hand, further study is necessary to elucidate the possible involvement of CK8 and CK14 in cadmium-induced carcinogenesis. It has been well documented that cancer cells have altered differentiation programs (keratin profiles) compared with their normal counterpart (12); however, this phenomenon has not been fully understood. In our study, we speculate that increased expression of specific keratins can act as a scaffold to bind to apoptotic factors, preventing them to do their functions, which can provide additional survival advantages to evade apoptosis. Taken together, our previous and current studies showed that cells responded to long-term cadmium exposure by using diverse antiapoptotic mechanisms (Fig. 4), thereby attenuating cadmium-induced apoptosis (9, 11). Our work also initiated ongoing research on the elucidation of the interactions between keratin and intracellular signaling as well as a better understanding of the mechanisms governing epithelial cell transformation.

Grant support: University of Hong Kong grants 10204004/39815/43700/301/01 and 10204565/38181/25200/301/01 and Research Grants Council grants HKU2718/02M and HKU7395/03M.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Celia N. L. Lau for critical reading of the manuscript and Y. Zhou, Anita Y.Y. Wong, and Cynthie Y.H. Cheung for technical assistance in our study at the University of Hong Kong.